The neuropathic phenotype of the K/BxN transgenic mouse with spontaneous arthritis: pain, nerve sprouting and joint remodeling

The adult K/BxN transgenic mouse develops spontaneous autoimmune arthritis with joint remodeling and profound bone loss. We report that both males and females display a severe sustained tactile allodynia which is reduced by gabapentin but not the potent cyclooxygenase inhibitor ketorolac. In dorsal horn, males and females show increased GFAP+ astrocytic cells; however, only males demonstrate an increase in Iba1+ microglia. In dorsal root ganglia (DRG), there is an increase in CGRP+, TH+, and Iba1+ (macrophage) labeling, but no increase in ATF3+ cells. At the ankle there is increased CGRP+, TH+, and GAP-43+ fiber synovial innervation. Thus, based on the changes in dorsal horn, DRG and peripheral innervation, we suggest that the adult K/BxN transgenic arthritic mice display a neuropathic phenotype, an assertion consistent with the analgesic pharmacology seen in this animal. These results indicate the relevance of this model to our understanding of the nociceptive processing which underlies the chronic pain state that evolves secondary to persistent joint inflammation.

Rheumatoid arthritis (RA) is characterized by chronic and symmetric inflammation of synovial joints leading to joint destruction, chronic pain, loss of function, and disability [1][2][3] . RA may also result in peri-articular bone erosions, and generalized osteoporosis with significant loss of bone mineral density (BMD) [4][5][6] . Current therapeutics, notably the biologics, have resulted in a significant improvement in the ability to manage the arthritic state [7][8][9] , but in spite of this efficacy in addressing the evolving joint dysfunction can, paradoxically, leave the associated pain unmanaged [10][11][12][13][14] . It is commonly thought that arthritic pain is mediated by inflammation or tissue damage, leading to a peripheral and central sensitization. Yet, the severity of the pain and the restricted efficacy of many anti-inflammatory agents often does not correlate well with these factors 15 . In fact, patients with RA frequently continue to report pain despite the control or resolution of clinical signs of arthritis 16 .
To better understand the complex mechanisms driving RA associated pain different animal models have been used that are largely induced and not spontaneously occurring [17][18][19][20] . Here we examine the K/BxN model 21 where mice express a transgenic T cell receptor, KRN, that drives the development of autoantibody formation to glucose 6 phosphate isomerase 22 . This autoantibody mediates the development of a robust autoimmune arthritis within a month after weaning in both male and female mice 21,23 . The structural characteristics of this arthritis display a phenotype similar to that found in human RA joints, including inflammatory infiltrates and articular erosions. Although the physical changes in bone and joint of the K/BxN mouse have been well described, the associated pain or the effects upon nociceptive afferent signaling have not been well characterized. Here we characterize in males and females the severity of arthritis, changes in bone parameters from the onset of mechanical allodynia.

Scientific RepoRtS
| (2020) 10:15596 | https://doi.org/10.1038/s41598-020-72441-5 www.nature.com/scientificreports/ Further our work and that of others has suggested the possibility that chronic inflammation may lead to changes in the innervation of the joint and activation of signaling in the dorsal horn and dorsal root ganglion (DRG) consistent with a neuropathic phenotype 17,19,20 . These changes together with analgesic pharmacology support the hypothesis that the K/BxN mouse displays a neuropathic phenotype which may be superimposed on the chronic inflammatory state.

Results
Adult male and female K/Bxn mice display severe erosive arthritis. We scored arthritis and measured mechanical withdrawal thresholds every 2 weeks after weaning up to an age of 16 weeks. All transgenic mice showed steadily increasing arthritis scores [F(2,127) = 221, p = 0.0001, two-way ANOVA; Fig. 1A]. At sixteen weeks male and female wild type and K/BxN mice were sacrificed and the hind limbs harvested and one limb was decalcified, and prepared for histology. The K/BxN mice had severe damage to the tibial/talar joint ( Fig. 1B 1H). There were no significant differences between sexes of each strain in mechanical withdrawal thresholds (Fig. 1H). Groups of 16-week-old mice were then tested for their responses to a nonsteroidal anti-inflammatory drug (NSAID), ketorolac given at a dose shown to robustly reduce inflammation and hyperalgesia in models of inflammation 24 , or gabapentin at doses which were efficacious in models of nerve injury and facilitated processing 25 (Fig. 1I) Male and female K/Bxn mice demonstrate trabecular bone loss. The hind limbs were also analyzed by microCT ( Fig. 2 and supplemental Table 1 for specific statistical analyses). We measured trabecular bone mineral density (tBMD), bone volume (BV/TV) and trabecular number (Tb.N) in the tibia ( Fig. 2A), calcanea (Fig. 2B) and tali (Fig. 2C). There were significant differences in bone structure between sex and strain ( Fig. 2 and supplemental Table 1 for specific statistical analyses). Between strains (WT versus K/BxN) there were significant differences in tBMD, BV/TV and Tb.N for both males and females in the tibia ( Fig. 2A), calcanea (Fig. 2B), and tali ( Fig. 2C) (p < 0.001, two-way ANOVA and Bonferroni post hoc test; Supplemental Table 1) with the exception of Tb.N in the tali of females (Fig. 2C). Similarly, there were significant differences between WT male and female mice in tBMD, BV/TV, and Tb.N of the  www.nature.com/scientificreports/ increases in sensory and tH + nerve fiber expression and sprouting in K/BxN mice. The ankle joints of 16-week-old K/BxN mice and controls were examined by immunohistochemistry for the quantification of sympathetic and peptidergic nerve fibers in the synovium. First, there was a profound inflammatory infiltrate seen by H&E staining in the joints of the transgenic mice, consistent with prior reports (Fig. 6A,B). In WT mice, a low-level, regular pattern of innervation by CGRP + and TH + nerve fibers was observed in the synovium.  www.nature.com/scientificreports/ In these mice, there was a robust increase in the density of TH + (used as marker for sympathetic nerves) 26,27 , CGRP + (sensory) and GAP-43 + fibers (nerve fibers undergoing sprouting) in both the male and female K/BxN mice ( Fig. 6C-K). In the transgenic mice, a significant number of CGRP + , TH + , and GAP-43 + nerve fibers had sprouted and had a disorganized appearance, as compared with wild type mice (Fig. 6). Quantitative analysis of the density of nerve fibers (expressed as density of nerve fibers) in the synovial membrane (ankle joint) revealed that the K/BxN group had a significant increase in the density of CGRP + fibers as compared with WT mice in males (819.9 ± 55.6 versus 67.0 ± 16.0 mm/mm 3  Whereas all K/BxN mice showed significant and robust increases in CGRP + , GAP-43 + , and TH + nerve fibers, indicative of sprouting, CGRP + fibers were structurally prominent in three out of nine male mice and one of eight female mice with at least 10 individual CGRP + axons that were > 20 µM thick (Fig. 7).

Discussion
In arthritic pathologies, such as RA, joints can remain painful after successful treatment of synovial inflammation. The quality of this pain can have neuropathic features 28,29 . One potential mechanism that may explain the dissociation between disease progression and pain in arthritis is an ectopic sprouting of sensory and sympathetic nerve fibers 18,19,30,31 . Here we examined male and female KRN T cell receptor transgenic mice for the onset of mechanical tactile allodynia and characterized them for central changes in the spinal cord, evidence of nerve injury in the DRG and peripheral sprouting of sympathetic and sensory nerves near the damaged joints. We evaluated both male and female animals, as there are reported differences in pain mechanisms between sexes [32][33][34] .
In this series of studies, clinical and histologic signs of inflammation did not display strong covariance with withdrawal thresholds. Several elements point to this characteristic as reflecting a neuropathic, as opposed to an inflammatory phenotype: www.nature.com/scientificreports/ (i) Ketorolac, a potent cyclooxygenase inhibitor 24 had a minimal impact on the mechanical allodynia. In contrast, gabapentin, which displays efficacy in neuropathic models through binding to the alpha 2 delta1 subunit of the voltage gated Ca 2+ channel (reviewed in 35 ), was highly effective. Importantly, while only a single dose of each was used, the ketorolac dose was chosen based on its efficacy in reducing inflammatory indices and hyperalgesia/ allodynia otherwise observed in robust models of inflammation 24 . Similarly gabapentin dosing was based on demonstrated efficacy and absence of effects upon spontaneous behavior in murine models of facilitated processing and polyneuropathy 25 . However, we note that there may have been time confounds in our protocol as the drugs were administered serially and not in a cross-over design.
(ii) In the adult K/BxN mice there was a noted sex difference in glial activation markers. Both males and females demonstrated an increase in GFAP but only males demonstrated an increase in Iba1 labeling which indicate activated astrocytes and microglia respectively. Glial activation has been implicated in the development and maintenance of a neuropathic condition 36 . Recent work has shown that microglia (Iba1) are not required for mechanical pain hypersensitivity in female mice. In one study, minocycline did not prevent mechanical allodynia female mice 32 . Moreover, only in male, pain related behavior seems to be driven by P2X4R-induced release of BDNF by microglia 37 .
(iii) Although ATF3 is frequently cited as a marker of nerve injury 38 , no increase in ATF3 staining was seen in the DRGs of these adult mice. We note that in the K/BxN and the collagen antibody-induced arthritis (CAIA) passive transfer models of arthritis, ATF3 + staining was demonstrated in the lumbar DRGs after approximately 10-15 days 17,20 . This increase in ATF3 was however typically transient and in the CAIA model the ATF3 staining was markedly diminished at a later time point 20 . Similarly, while nerve ligations also lead to significant increases in DRG-ATF3 expression, this increase secondary to direct nerve injury can disappear by 14 days 39 . Hence, the absence of DRG-ATF3 staining was not inconsistent with a neuropathic phenotype. As the present immunohistochemical studies focused on adults, it is reasonable that we might find such changes during an earlier stage of development in this arthritis phenotype.
(iv) In the K/BxN transgenic mice, CGRP + , TH + , and GAP-43 + nerve fibers were observed to show a significant increase in fiber density and, in contrast to the regular pattern of innervation seen in WT mice, displayed a www.nature.com/scientificreports/ highly ramified disorganized appearance. CGRP is uniformly considered to reflect small afferent innervation 40 . TH has been extensively used as marker for sympathetic post-ganglionic innervation to study peripheral sprouting in pathological conditions 18,27 . However, we note that TH is also expressed in a subpopulation of small DRG neurons 41 . This raises the question whether the peripheral joint innervation expressing TH reported here represents sensory, post ganglionic sympathetic or both. TH expressing DRG neurons innervating non-visceral tissues seldom co-express CGRP, in contrast to visceral innervation 42 . It thus appears likely that at least some of the upregulated expressing axons represent post ganglionic fibers. Additional work is required to address this possibility. In all, these findings are consistent with previous studies in humans and rodents indicating that remodeling of sensory and sympathetic nerve fibers can be associated with musculoskeletal pain 18,19,30,31,43,44 . Sprouting of afferents and TH + fibers along with the formation of neuroma like structures can lead to spontaneous discharges [45][46][47] . This property of initiating ectopic/spontaneous discharges can explain in part components of the chronic arthritis pain, which is refractory to anti-inflammatory interventions. It should be noted that this sprouting may be a generalized response to chronic joint inflammation. Significant increases in CGRP, GAP-43 and TH have also been reported after repeated intra articular injections of complete Freund's adjuvant 18,31 . Hence, the ectopic sprouting of sensory and sympathetic nerve fibers reported in separate models of chronic murine arthritis, suggests a common mechanism may be involved in the generation and maintenance of the arthritic pain in the chronic animals. These observations suggest that the transgenic K/BxN mouse provides a specific phenotype for the chronic inflammatory state that may arise from an early onset joint inflammation and suggest that that underlying pain state reflects a polyneuropathic phenotype in both sexes. This distinction may be of particular significance in screening for the development of new analgesic therapeutics for these chronic states given the robust distinction observed here between the actions of a NSAID and a centrally active antihyperalgesic, gabapentin. Arthritis scoring. The development of joint inflammation in the paws was evaluated by visual inspection and rated on a scale of 1 to 28 where one point was given for each swollen digit and two points for each swollen ankle or wrist (modified from a prior report 48 ).

Methods
Von frey behavioral testing. Mechanical withdrawal thresholds were assessed every other week using the up-down method 49 . Briefly, animals were placed in clear, plastic, wire mesh-bottomed cages for 45-min prior to the initiation of testing. Tactile thresholds were measured with a series of calibrated von Frey filaments (Seemes Weinstein von Frey anesthesiometer; Stoelting Co., Wood Dale, IL, USA) ranging from 2.44-4.31 (0.02-2.00 g). Mechanical values for the left and right hind paws were measured and the 50% probability of withdrawal threshold was calculated 49 . Thresholds for the two hind paws were averaged to produce a single data point per day of measurement.
Drugs and drug delivery. To define the analgesic pharmacology of the K/BxN allodynic state, groups of 16-week-old K/BxN transgenic arthritic mice (n = 6/sex) received a single i.p. injection of gabapentin (100 mg/ kg; Pfizer), or the NSAID, ketorolac (7.5 mg/kg; Sigma). Mice were measured for changes in tactile allodynia 1 h following treatment. Drugs were allowed to wash out for 48 h before the same procedure was repeated with a different drug. Gabapentin 50,51 , and ketorolac 24 were dissolved in normal saline, dosages were based on previous reports demonstrating anti-allodynic efficacy, absent motor or general behavioral effects, in murine models of mono / poly and inflammatory neuropathy, respectively.
Histology and immunohistochemistry. At sixteen weeks, mice were deeply anesthetized with Beuthanasia-D and perfused intracardially with 0.9% saline followed by 4% paraformaldehyde. The spinal cords and lumbar DRGs were removed, post-fixed, and cryoprotected in 30% sucrose. The ankle in the mouse is innervated with varying contributions by levels L3-L5 depending on the strain, with C57BL/6 mice trending toward L4-L5 52  www.nature.com/scientificreports/ cyte (GFAP) labeling was quantified by measuring the fluorescence intensity converted to optical density (OD) in arbitrary units in laminae I and II of the dorsal horn of the spinal cord in ImageJ. Initially each picture was converted to 8-bit gray scale. We then calibrated each section using the "uncalibrated OD" function. Four circular regions of interest (diameter = 88 µM) were placed in laminae I and II. The OD from an equivalent area in a deep layer of the same slice was subtracted as the background from the average reading. The L4-L5 DRGs were cut in transverse sections (14 μm) and mounted on glass slides. DRGs were incubated with anti-NeuN (neuronal nuclei) biotinylated (1:1,000, Chemicon), anti-ATF3 (1:1,000, Santa Cruz), anti-Iba1 antibody (1:1,000 Wako, Richmond, VA), anti-TH (1:1,000, Millipore), anti-CGRP (1:1,000, Sigma) and nuclear stain 4′,6-diamidino-2-phenylindole (DAPI). Binding sites were visualized with secondary antibodies conjugated with fluoro-Alexa-488. The total number of immunoreactive cells were counted in each section (four sections per animal). The number of NeuN positive nuclei per section was used as the denominator for the percentage of neurons that were CGRP + , TH + or ATF3 + . At least six animals are included for each group.
To analyze periarticular innervation, once hind extremities were harvested, they were post-fixed for 24 h and placed in 0.01 M PBS (pH = 7.4) at 4 °C until their micro computed tomography (microCT) analysis. Once microCT was finalized, samples were placed in a 10% EDTA solution for two weeks for decalcification. The degree of decalcification was monitored at regular intervals by plain radiography. Then, decalcified bones were cryoprotected in a 30% sucrose solution at 4 °C for further processing by immunohistochemistry. The femora were cut into serial cross sections (20 μm thick) with a cryostat (MTC, SLEE, Mainz, Germany) 31,55-57 . Frozen ankle joint sections were incubated with antibody against calcitonin gene-related peptide (CGRP, polyclonal rabbit anti-rat CGRP; 1:4,000; Sigma) to label primary afferent sensory nerve fibers. Sympathetic nerve fibers were labeled with antibody against tyrosine hydroxylase (TH, polyclonal rabbit anti-rat TH, 1:1,000; EMD Millipore). Sprouted nerve fibers were labeled with an antibody against growth-associated protein-43 (GAP-43, rabbit anti-GAP-43, 1:1,000; EMD Millipore). Nuclear stain 4′,6-diamidino-2-phenylindole (DAPI) was used to visualize all cell bodies.
For quantification of nerve fiber density, approximately 15 separate, 20-μm-thick frozen sections were obtained from the ankle joint of each mouse. For each given marker, three sections from one ankle joint were initially scanned at low magnification (× 10) to identify the areas with the highest density of nerve fibers thorough an epifluorescence microscope (Axio Scope.A1, Carl Zeiss, Jena, Germany). These areas in the synovium were defined as "hot spots" 31 . One image per section was acquired within the medial synovial hot spot. Then, an image was obtained for each section in this area (separated by at least 100 μm) at 40 × magnification with an epifluorescence microscope. The images were analyzed using ImageJ software (National Institutes of Health) and nerve fibers were manually traced by a blinded investigator, using the freehand line tool, to determine the total length of nerve fibers. Data from at least three sections per ankle joint were recorded and averaged. Total volume of the synovium (naïve and inflamed) was calculated by tracing the area of the synovium and multiplying this area by the thickness of the section (20 μm). Data are expressed as total length of nerve fibers per volume of synovium (mm/mm 3 ) 18,26,31 . Representative images from WT and transgenic mice for each marker were captured with a Carl Zeiss scanning confocal laser microscope (Model LSM 800, Jena Germany). Confocal images used for illustration were assembled and labeled using Adobe Illustrator software. At least two sections (10-μm-thick) from each mouse, were stained with hematoxylin and eosin (H&E) to identify the morphology and integrity of the ankle joints. Representative images from H&E-stained sections were captured with a light microscope Axio Scope.A1, Carl Zeiss, Jena, Germany). One section per mouse was scored in a blinded manner for inflammation and erosion using a semi-quantitative scale (0-4) as previously described 58 . Microcomputed tomography. To assess changes in bone parameters, microCT analysis was performed in the trabecular tibia, calcaneus, and talus using a desktop microCT system (Skyscan 1272; Bruker, Belgium). The scanning parameters were: 10 µm voxel size, at 60 kVp and 166 µA with an integration time of 627 ms, according to the guidelines for microCT analysis of rodent bone structure 59 . All the scanner images were reconstructed using NRecon Software (Bruker, Belgium). The trabecular region of interest (ROI) at distal femur and proximal tibia was evaluated by selecting 2 mm in the vertical axis, subsequent to 0.5 mm from the growth plate (reference point). For the calcaneus, analysis ROI was selected using a 0.5 mm 2 cylinder that was positioned underneath the growth plate of the calcaneus bone and finally for the talus analysis ROI was selectively bordered manually. The CT analyzer program (Bruker, Belgium) was used to determine trabecular bone parameters; an automatic segmentation algorithm (CT analyzer) was applied in order to isolate the trabecular bone from the cortical bone. The parameters used for the trabecular bone were: trabecular bone mineral density (BMD), trabecular bone volume rate (BV/TV), and trabecular number (Tb.N). Finally, hydroxyapatite calibration phantoms (250 and 750 mg/cm 3 ) were used to calibrate trabecular bone density values (BMD).

Statistics.
Results are represented as means ± SEM. Statistical analysis was performed using GraphPad Prism (version 8.3.0; GraphPad Software, San Diego, CA, USA). Minimum group sizes were predicated on extensive previous work 17,60,61 and demonstrated as appropriate for establishing as statistically significant (p < 0.05) changes in paw diameter and tactile thresholds that were considered to be biologically and behaviorally relevant, as compared to vehicle or baseline controls. This group size was also considered to be relevant for demonstrating joint sprouting and arthritic bone remodeling based on our previous work 62 . For multiple comparisons one or two-way ANOVA was used with Bonferroni post hoc test. In all cases, p < 0.05 was considered significant.

Data availability
Authors can confirm that all relevant data are included in the article and/or its supplementary information file.